Merge pull request pandas-dev#6447 from cpcloud/read-html-float-iterable-5129

BUG/TST: read_html should follow pandas conventions when creating empty data

Author: cpcloud

doc/source/release.rst

@@ -184,6 +184,11 @@ Bug Fixes
 - Bug in ``sum`` of a ``timedelta64[ns]`` series (:issue:`6462`)
 - Bug in ``resample`` with a timezone and certain offsets (:issue:`6397`)
 - Bug in ``iat/iloc`` with duplicate indices on a Series (:issue:`6493`)
+- Bug in ``read_html`` where nan's were incorrectly being used to indicate
+  missing values in text. Should use the empty string for consistency with the
+  rest of pandas (:issue:`5129`).
+- Bug in ``read_html`` tests where redirected invalid URLs would make one test
+  fail (:issue:`6445`).
 
 pandas 0.13.1
 -------------

pandas/io/html.py

@@ -579,8 +579,9 @@ def _expand_elements(body):
     lens_max = lens.max()
     not_max = lens[lens != lens_max]
 
+    empty = ['']
     for ind, length in iteritems(not_max):
-        body[ind] += [np.nan] * (lens_max - length)
+        body[ind] += empty * (lens_max - length)
 
 
 def _data_to_frame(data, header, index_col, skiprows, infer_types,
@@ -760,15 +761,15 @@ def read_html(io, match='.+', flavor=None, header=None, index_col=None,
         the table in the HTML. These are not checked for validity before being
         passed to lxml or Beautiful Soup. However, these attributes must be
         valid HTML table attributes to work correctly. For example, ::
-        
+
             attrs = {'id': 'table'}
-        
+
         is a valid attribute dictionary because the 'id' HTML tag attribute is
         a valid HTML attribute for *any* HTML tag as per `this document
         <http://www.w3.org/TR/html-markup/global-attributes.html>`__. ::
-        
+
             attrs = {'asdf': 'table'}
-        
+
         is *not* a valid attribute dictionary because 'asdf' is not a valid
         HTML attribute even if it is a valid XML attribute.  Valid HTML 4.01
         table attributes can be found `here